Download Relative importance of endogenous and exogenous mechanisms in

16 (4): 429-440 (2009)
Relative importance of endogenous and exogenous
mechanisms in maintaining phytoplankton
species diversity1
Anita NARWANI2, Julie BERTHIN & Asit MAZUMDER, Water and Aquatic Sciences Research Program,
Department of Biology, University of Victoria, P.O. Box 3020, Station C.S.C., Victoria, British Columbia
V8W 3N5, Canada, e-mail: [email protected]
Abstract: The competitive exclusion principle poses the pressing question of how biodiversity is maintained in nature. Many
mechanisms have been proposed to explain diversity and to resolve what has become known as the “paradox of the plankton”.
We propose a dichotomy among these mechanisms in order to enable empiricists to begin testing their relative importance.
Specifically, the mechanisms can be categorized as being internally generated or as depending on forces external to the
competitive community. Here we tested whether the internal competitive dynamics of a phytoplankton assemblage or
externally generated resource variability (a disturbance) were more effective at maintaining species diversity over time. We
also tested whether the species composition of assemblages was important in determining the persistence of species diversity.
We employed controlled microcosm experiments in which we either imposed exogenous variability in nutrient availability via
serial dilution or allowed the communities to remain completely undisturbed. We found that species diversity was maintained
most effectively in undisturbed microcosms in which only internal dynamics regulated coexistence. We also found that
the community composition of the assemblage significantly interacted with the disturbance regime in determining species
diversity. This confirmed the importance of internal dynamics and community composition in maintaining species diversity.
Keywords: coexistence, community composition, competition, disturbance, species diversity.
Résumé : Comment la biodiversité est maintenue dans la nature est une question fondamentale découlant du principe
d’exclusion compétitive. De nombreux mécanismes ont été proposés pour expliquer la diversité et résoudre ce qui est connu
comme le « paradoxe du plancton ». Nous proposons une dichotomie parmi ces mécanismes pour permettre aux empiristes
de commencer à évaluer leurs importances relatives. Ainsi, les mécanismes peuvent être divisés comme étant soient générés
de l’intérieur ou causés par des forces externes à la communauté en compétition. Nous avons testé ici si la dynamique
compétitive interne d’un assemblage de phytoplanctons ou un facteur externe, la variabilité des ressources (une perturbation)
était plus efficace à maintenir la diversité d’espèces dans le temps. Nous avons aussi évalué si la composition en espèces d’un
assemblage était importante pour la persistance de la diversité d’espèces. Nous avons réalisé des expériences en microcosmes
contrôlés dans lesquelles nous avons soit imposé de la variabilité exogène dans la disponibilité en nutriments par des dilutions
successives ou gardé les communautés intactes, sans aucune perturbation. Nous avons constaté que la diversité d’espèces
était maintenue plus efficacement dans les microcosmes non perturbés dans lesquels seule la dynamique interne était
responsable de la coexistence. Nous avons aussi constaté que la composition de la communauté d’un assemblage et le régime
de perturbation interagissaient de façon significative afin de déterminer la diversité d’espèces. Cela a confirmé l’importance
de la dynamique interne et de la composition de la communauté dans le maintien de la diversité d’espèces.
Mots-clés : coexistence, compétition, composition de la communauté, diversité d’espèces, perturbation.
Introduction
The competitive exclusion principle states that only a
single competitor can survive on a single limiting resource
(Volterra, 1928; Hardin, 1960). The principle has been
extended to show that n species require at least n resources
(MacArthur & Levins, 1964), “limiting factors” (Levin,
1970), or niches (Rescigno & Richardson, 1965) to ensure
indefinite and stable equilibrium coexistence in a homogeneous environment. The principle has inspired a search for
mechanisms that maintain species diversity in nature despite
competition’s tendency to reduce it, given the generally
smaller number of limiting factors identified in most eco-
1Rec.
2008-10-23; acc. 2009-08-18.
Associate Editor: Suzanne Roy.
2Author for correspondence.
DOI 10.2980/16-4-3232
systems (Hutchinson, 1961; Connell, 1978). These mechanisms have included limits to similarity and niche differentiation (MacArthur & Levins, 1967; MacArthur, 1969; Tilman,
1982; Chase & Leibold, 2003), disturbance of equilibrium
species dynamics (Connell, 1978; Huston, 1979), densityindependent mortality (Koch, 1974a; Abrams, 2001; Steiner,
2005), density- and frequency-dependent mortality (Janzen,
1970; Connell, 1978; Holt, 1984), predation (Caswell, 1978;
Crowley, 1979; Holt, 1984), heterogeneity in the environment over space or time (Chesson & Huntly, 1997; Chesson,
2000a), storage effects (Chesson & Warner, 1981; Chesson,
2000a), competition–colonization trade-offs (Levins &
Culver, 1971; Tilman, 1994), and species equivalence or
neutrality (Hubbell, 2001; Scheffer & Van Nes, 2006). This
list incorporates a great number of non-exclusive mechanisms, and there have been efforts to unite them all under a
common synthetic framework.
NARWANI, BERTHIN & MAZUMDER: ENDOGENOUS DYNAMICS MAINTAIN SPECIES DIVERSITY
Such a framework was most recently proposed by
Chesson (2000b), who showed that mechanisms maintaining coexistence can work in 2 distinct ways: either
they minimize fitness differences between species (these
are equalizing mechanisms) or they increase the relative
strength of intraspecific limitation over interspecific limitation (stabilizing mechanisms). This framework employs
a modern definition of stable coexistence, which can be
broadly equated with long-term persistence (DeAngelis &
Waterhouse, 1987). Such a view of stable coexistence is
fundamentally different from that held by Volterra (1928)
and Gause (1934) because it does not hinge on the existence
of a stable equilibrium point, and stabilizing mechanisms
may depend upon population fluctuations. As a result, competition is not necessarily a destructive force, pushing the
system towards dominance by a single species; when competition is relatively nonlinear between species, it too can
induce fluctuations and stabilize coexistence (Koch, 1974b;
Armstrong & McGehee, 1980; Huisman & Weissing, 1999;
Abrams & Holt, 2002).
In this paper we propose that, given Chesson’s (2000b)
new framework, recognition of a second major dichotomy
among mechanisms will aid empiricists in comparing their
relative importance. Namely, mechanisms either depend
on exogenous forces or structure (e.g., variability in limiting factors or spatial structure) or depend on internal
competitive dynamics that can result in increased resource
limitation or variability in limiting factors (Tilman, 1982;
Chesson, 2000b). Exogenous variability in limiting factors
can allow the expression of storage effects and relative nonlinearities in competitive ability among species (Chesson,
2000b). Density-independent mortality, disturbances,
nutrient pulses, heterogeneity in resource supply, densitydependent specialist predation, and frequency-dependent
predation all represent sources of environmental variability
that are extrinsic to the assemblage of competitors and may
enable coexistence. Intrinsic competitive dynamics, on the
other hand, may also promote non-equilibrium coexistence
when competitive abilities are relatively nonlinear and produce intrinsic fluctuations in limiting factors (Armstrong
& McGehee, 1980). When the limiting factors are abiotic
resources, competition among 3 or more species can produce “supersaturated” coexistence, or the coexistence of > n
species on n resources. The likelihood of such coexistence
has been debated (Schippers et al., 2001). However, the
probability of supersaturated coexistence can be high (up to
~70%) when competitors display intransitive competition.
Intransitive competition occurs when 3 or more competitors
form competitive networks rather than hierarchies, as in the
common children’s game of rock–paper–scissors (Huisman
& Weissing, 2001). In this scenario, no competitor is superior to all other competitors, but each competitor is superior
to some, e.g., rock beats scissors, scissors beat paper, paper
beats rock. Competitive intransitivity has been shown to
exist and to enable coexistence in bacterial (Kerr et al.,
2002; Kirkup & Riley, 2004) and reef invertebrate communities (Buss & Jackson, 1979), as well as among morphs
of male lizard within a population (Sinervo & Lively, 1994).
The relative importance of exogenous and endogenous
mechanisms of coexistence in maintaining biodiversity is
430
unknown. Recently, temporal storage dynamics, in which
coexistence is promoted by external environmental variability, have been demonstrated both in the lab (DescampsJulien & Gonzalez, 2005) and in the field (Kelly & Bowler,
2002), although spatial storage dynamics are predicted
to occur more commonly (Chesson, 2000a). Sources of
exogenous variability such as disturbance and predation do
not promote coexistence simply by reducing the strength
of competition (Chesson & Huntly, 1997), but rather they
do so also by allowing the expression of relative nonlinearity in competition and storage dynamics (Chesson, 2000a).
By contrast, with regard to internal dynamics, there have
been numerous demonstrations of competitive exclusion in
very simple, purely competitive systems (e.g., Gause, 1934;
Tilman, 1977; 1981). Nevertheless, some more recent studies have shown that numerous species can coexist apparently indefinitely in purely competitive communities that
are not subject to exogenous variability in limiting factors
(Kerr et al., 2002; Roelke, Augustine & Buyukates, 2003;
Haddad et al., 2008). Our main aim here was to test the
relative abilities of internally and externally driven dynamics to maintain species diversity.
Theory also predicts that the coexistence of multiple
species should depend upon the traits of the species in the
assemblage (Huisman & Weissing, 2001; Chase & Leibold,
2003). For instance, niche theory states that, at equilibrium,
species with similar requirements but different impacts
on their limiting factors will be more likely to coexist
(Chase & Leibold, 2003). Away from equilibrium, relative nonlinearity in competitive abilities or storage effects
are required, both of which entail particular differences
in species responses to limiting factors (Chesson, 2000b).
Huisman and Weissing (2001) showed that supersaturated
coexistence is more likely when species consume the most
of the resource for which they have intermediate requirements. Species traits are therefore predicted to be important
in determining the ability of species to coexist under both
equilibrium and nonequilibrium conditions and regardless of whether variability in limiting factors is generated
internally or externally. Thus, we also tested the effects of
species identity and community composition on coexistence
over time.
Specifically we addressed the following questions: 1) Is
exogenous variability in limiting factors or endogenously
generated competitive dynamics more effective at maintaining species diversity in assemblages of phytoplankton?
2) Does the maintenance of species diversity depend on the
species composition of the experimental community?
Methods
We tested the effects of external variability on species
diversity by imposing density-independent mortality and
fluctuations in resource availability on the communities.
We compared the species diversity maintained in these
treatments over time to that maintained in completely undisturbed, control treatments. We crossed these “disturbance”
levels by 3 different community compositions in order to
determine the importance of the identity of species within
communities on coexistence.
ÉCOSCIENCE, VOL. 16 (4), 2009
We created 3 unique phytoplankton community compositions by randomly assigning 4 species to each composition from a total pool of 9 species (Table I). In order to
measure the growth rate of each species under our experimental conditions, we recorded the raw fluorescence (in raw
fluorescence units, or RFUs) of well-mixed samples of each
species daily for 18 d after an initial inoculation of batch
monocultures. We measured raw fluorescence of each species with a Turner Designs Trilogy Laboratory Fluorometer,
(Turner Designs Inc., Sunnyvale, California, USA) and calculated their intrinsic maximal growth rates, μmax (d–1):
( )
P
μ max = 1 ln t
T
Po
[1]
where Po = RFUs at the beginning of exponential growth,
Pt = RFUs at the end of exponential growth, and T = days
spent in exponential growth (Kilham et al., 1998).
The microcosms (500 mL Erlenmeyer flasks) each
initially contained 250 mL of COMBO freshwater culture
medium (Kilham et al., 1998). The phosphate and nitrate concentrations in COMBO are 50 μmol·L–1 and 1000 μmol·L–1,
respectively (Kilham et al., 1998). We inoculated equal cell,
colony or filament densities into the experimental microcosms using sterile technique. For the inoculations, we used
exponentially growing monocultures of phytoplankton,
which were maintained in serially diluted batch culture. We
measured the densities of each of the monocultures using a
Bright-Line hemacytometer (Hausser Scientific, Horsham,
Pennsylvania, USA) and either concentrated or diluted
them to obtain suspensions with equal densities of ~30 000
cells·mL–1. To inoculate, we removed 20 mL of medium from
each microcosm and replaced it with 5 mL of monoculture
from each of 4 species using sterile technique. After inoculation, the assemblages grew in an Enconair Environmental
Growth Chamber (Econair Environmental Growth Chambers
Inc., Winnipeg, Manitoba, Canada) at a constant temperature of 20 °C (max/min deviation of 1 °C). They received
79.7 ± 7.09 (mean ± SD) μmol photons·m–2·s–1 of photosynthetically active radiation from cool-white fluorescent lights
on a 12 h light:12 h dark cycle. Throughout the experiment,
the microcosms were shaken only before imposing disturbTABLE I. The 4 species introduced to each community composition.
Inocula of phytoplankton species were obtained from the University
of Texas Culture Collection (UTEX) and the Canadian Phycological
Culture Centre at the University of Waterloo (CPCC).
Community
composition
Species (Source and identifier)
1
Ankistrodesmus falcatus (UTEX 101)
Chlamydomonas reinhardtii (CPCC 84)
Pseudokirchneriella subcapitata (CPCC 37)
Staurastrum pingue (UTEX 1606)
Rhodomonas minuta (CPCC 344)
Cryptomonas cf. rostratiformis (CPCC 343)
Ankistrodesmus falcatus (UTEX 101)
Fragilaria crotonensis (CPCC 269)
Asterionella formosa (CPCC 605)
Fragilaria crotonensis (CPCC 269)
Cyclotella sp. (CPCC 537)
Cryptomonas cf. rostratiformis (CPCC 343)
2
3
Growth rate
(d–1)
0.513
0.778
0.281
0.204
0.197
0.201
0.513
0.425
0.231
0.425
0.703
0.201
ances and before sampling to ensure that representative
communities were sampled.
Three community compositions were crossed by 3
disturbance levels in a factorial design. The 3 disturbance
levels consisted of serial dilutions at 7-d intervals, serial
dilutions at 11-d intervals, and a treatment in which the
microcosms were undisturbed for the duration of the entire
experiment (97 d). The disturbed microcosms experienced
density-independent death and nutrient replenishment. The
undisturbed microcosms did not receive any nutrient replenishment for the full duration of the experiment. We imposed
disturbance treatments by shaking the microcosms and then
diluting the phytoplankton assemblages with autoclavesterilized medium. Specifically, this involved the removal of
150 mL of the phytoplankton assemblage from the shaken
Erlenmeyer flasks and replacement of this volume by fully
enriched sterile COMBO. We chose the duration of the 2
disturbance intervals because previous disturbance experiments have shown that peaks in diversity tend to occur at
intervals between 5 and 28 d, with most peaks occurring
at intervals between 7 and 10 d (Robinson & Sandgren,
1983; Gaedeke & Sommer, 1986; Sommer, 1995; Flöder
& Sommer, 1999; Flöder, Urabe & Kawabata, 2002). We
thus hoped to maximize the likelihood of detecting storage
effects expressed as a result of fluctuating nutrient availability. Each community composition by disturbance treatment
cross was replicated twice.
We sampled 15 mL of each experimental microcosm
after shaking at low speed for 1 min. We did this every 2
weeks for the first month, and then at 1-month intervals
until 97 d had passed. We fixed the 15-mL samples with
150 μL (1% by volume) of Lugol’s iodine solution. We then
concentrated the samples to 4 mL by allowing them to settle
for 10 h and removing 11 mL of supernatant. We counted
species-specific cell densities in the concentrated samples
with a Bright-Line hemacytometer. For species with densities > 100 000 cells·mL–1, we counted only the middle
quadrat, with a volume of 0.1 μL. Otherwise, we counted
the number of cells in the entire counting chamber, with a
volume of 0.9 μL. The counting procedure was replicated
6 times for each sample. We did not replace the sample
volume removed from the microcosms with fresh medium
because this would have introduced a pulse of nutrients and
would therefore have interfered with the disturbance treatments. All of the microcosms therefore decreased in volume
by a total of 75 mL over the course of the experiment. Thus,
the intensity of the disturbances increased after each sampling interval from a reduction in density of 60% to 63.8%
to 68.2% to 73.0% and finally to 78.9%.
On the final sampling date we took samples from
each microcosm for organic carbon and nutrient analyses
to determine the level of resource limitation. We measured
total organic carbon, total phosphorus, and total nitrogen
from unfiltered samples. We calculated the particulate
portion of carbon and nutrients as the total concentration
minus the dissolved fraction. We filtered samples through
Whatman GF/F glass microfiber filters (Whatman PLC,
Maidstone, Kent, UK) to measure the dissolved organic carbon in the filtrate. We immediately ran these samples using
a Shimadzu TOC-V-CPH Total Organic Carbon Analyzer
431
NARWANI, BERTHIN & MAZUMDER: ENDOGENOUS DYNAMICS MAINTAIN SPECIES DIVERSITY
(Shimadzu Corporation, Kyoto, Japan). We used Fisherbrand
0.45-μm nitrocellulose membrane filters (Fisher Scientific
Ltd., Ottawa, Ontario, Canada) when filtering for the analysis
of dissolved nitrogen and phosphorus, and we froze these
samples for analysis within 4 weeks of sampling. We measured phosphorus as orthophosphate (PO43–) by the molybdenum blue–ascorbic acid method (Clesceri, Greenberg
& Eaton, 1998). We measured inorganic nitrogen using
cadmium-reduction of nitrate to nitrite, followed by quantification of nitrite according to the sulphanilamide method
(Clesceri, Greenberg & Eaton, 1998).
We used cellular biovolume as a measure of biomass.
We calculated the biovolume of each species by measuring
the relevant cellular dimensions and plugging these into previously described geometric formulae for their cell volumes
(Hillebrand et al., 1999). We multiplied cell counts by cell
biovolumes to achieve population biovolumes.
We used the Shannon Index as a measure of diversity
(Shannon & Weaver, 1949):
H' = Σ pi·ln pi
[2]
where pi is the relative abundance of species i. We derived pi
Bi
from biovolume, B, as
.
Btotal
We used biovolume (instead of abundance) in the
Shannon Index so that the weighting of each species on the
index would better reflect their ecological impacts (e.g.,
depletion of nutrients or ability to support consumers).
We calculated evenness as
E=
H'
Hmax
[3]
as a covariate, with disturbance as a fixed effect. We also
tried this using the ratio of particulate organic carbon to
particulate nitrogen (POC:PN) because it appeared that
nitrogen may have been more limiting than phosphorus
(Appendix II). We compared these results to the ANOVA
without the covariate to determine whether the inclusion
of the covariate eliminated disturbance from the model as
a significant explanatory variable. We ln-transformed H´
to meet the assumption of homogeneity of variances. We
used a Dunnett’s post hoc comparison to test for differences
between the disturbed treatments (7-d and 11-d intervals)
and the undisturbed treatment.
We used ANOVA to determine whether disturbance
level had an effect on particulate organic carbon (POC). In
order to meet the assumption of homogeneity of variances,
we ln-transformed the POC data. We used a Tukey post hoc
comparison to determine which disturbance level(s), if any,
differed from the others. We also used ANOVA to determine
whether disturbance had an effect on the carbon:nitrogen
(C:N) and carbon:phosphorus (C:P) molar ratios. We lntransformed the C:P ratios to meet the assumption of homogeneity of variances. We used SPSS v 15.0 for all numerical
analyses (SPSS Incorporated, Chicago, Illinois, USA).
Results
The highest mean species richness (S), diversity (H´,
Figure 1), and evenness (E) occurred in the undisturbed
microcosms when we considered community compositions
as replicates (Table II). The lowest values occurred under
the 7-d disturbance treatment. The same pattern was evident
when we compared the effects of disturbance on diversity
within each community composition (Figure 2).
where H’ is the Shannon Diversity Index, and H´max is the
maximum value of H’, and
H'max = ln S
[4]
where S represents species richness.
We tested the effects of disturbance, community composition, and their interactions over time on phytoplankton diversity (H´) using a repeated measures analysis of
variance (rm-ANOVA). In order to derive estimates for 3
missing data points, we used linear interpolation between
bracketing data. In 2 cases, a missing value occurred at the
beginning of a time series, in which case we used the subsequent values. We found that the effects of disturbance and
community composition interacted with time (Table IIIb),
which made the interpretation of their main effects questionable (Underwood, 1997). As a result, we compared only
the data on the final sampling dates (day 97), in a factorial
analysis of variance (ANOVA).
We found that greater nutrient limitation occurred in
the undisturbed microcosms by the end of the experiment
(see Results), and we wanted to test whether the effects of
disturbance on diversity could be explained by the degree
of nutrient limitation. To do this, we used the ratio of particulate organic carbon to particulate phosphorus (POC:PP)
as a measure of nutrient limitation (Hecky & Kilham,
1988), and we entered this into a single-factor ANOVA
432
FIGURE 1. The Shannon Diversity Index (H´) as a function of the disturbance interval. Each point represents the mean H´ of each disturbance treatment on the final sampling day. Bars represent one standard error (n = 3).
ÉCOSCIENCE, VOL. 16 (4), 2009
TABLE II. Mean species diversity, evenness, richness, and carbon:nitrogen:phosphorus ratio (C:N:P) of each disturbance treatment at the
termination of the experiment (t = 97 days). SE are shown in parentheses (n = 6).
Disturbance treatment
7-d
11-d
Undisturbed
Shannon Diversity Index (H')
Evenness (E)
Species richness (S)
C:N:P
0.026 (0.011)
0.379 (0.091)
0.889 (0.158)
0.028 (0.013)
0.346 (0.082)
0.618 (0.087)
2.333 (0.333)
2.833 (0.167)
4.167 (0.307)
32:11:01
50:07:01
131:13:01
TABLE III. a) Results of the repeated measures-ANOVA betweensubjects effects on species diversity, measured as the Shannon Index
(H'). b) Repeated measures-ANOVA within-subject effects. c) Results
of the factorial ANOVA examining the effects of disturbance, community composition, and their interactions on H' for day 97. d) Results of the ANOVA examining the effect of disturbance on H' for
day 97. e) As for d, except POC:PP was included in the model as a
covariate. CC = community composition, Dist = disturbance.
Source
FIGURE 2. An interaction plot of the effects of disturbance and community composition on the diversity (H´) of the microcosms on day 97. Circles
represent community composition 1, triangles represent composition 2, and
squares represent composition 3. Error bars represent ± 1 SE of the mean
for each composition × disturbance treatment cross (n = 2).
The rm-ANOVA revealed that community composition
(P = 0.001), disturbance (P < 0.001) and their interaction
(P = 0.009) all had significant between-subjects effects
on H´ (Table IIIa; Figure 3). The within-subjects effects of
disturbance and community composition on H´ interacted
significantly with time (Table IIIb). As a result, the main
effects of disturbance and community composition could
not be clearly interpreted (Underwood, 1997). The factorial ANOVA for the effects of disturbance and community
composition on H´ on day 97 showed that disturbance and
its interaction with community composition were significant (P < 0.001 and P = 0.004, respectively, Table IIIc;
Figure 2). Increasing disturbance interval length caused
an increase in diversity for all of the community compositions. However, the strength of the increase between
consecutive disturbance levels varied among community
compositions (Figure 2). The increase in H´ between the
7-d and 11-d interval was weaker for community composition 1 than for compositions 2 and 3, but it was stronger
for composition 1 than for compositions 2 and 3 between
the 11-d and undisturbed treatments.
Because the disturbance treatments had the same qualitative effects (i.e., in terms of direction, though not in magnitude) on all of the community compositions, we judged it
Type III sum Degrees of
of squares
freedom
a)
CC
1.482
Dist
2.409
CC × Dist
1.014
Error
0.349
b)
Time
1.361
Time × CC
1.739
Time × Dist
1.650
T × CC × Dist
1.470
Error (Time)
1.356
c)
R2 = 0.971
CC
0.061
Dist
2.260
CC × Dist
0.750
d)
R2 = 0.864
Dist
66.248
Error
10.451
e)
R2 = 0.865
Dist
60.205
POC:PP
0.099
Error
10.353
Mean
square
F
P
2
2
4
9
0.741
1.205
0.253
0.039
19.129
31.097
6.542
0.001
<0.001
0.009
4
8
8
16
36
0.340
0.217
0.206
0.092
0.038
9.034
5.770
5.476
2.439
<0.001
<0.001
<0.001
0.013
2
2
4
0.031
1.130
0.188
1.439
52.964
8.791
0.287
<0.001
0.004
3
6
22.083
1.742
12.677
0.005
3
1
5
20.068
0.099
2.071
9.692
0.048
0.016
0.836
to be meaningful to interpret the main effect of disturbance
on diversity, despite the significant interaction with community composition (Quinn & Keough, 2002). The singlefactor ANOVA showed that disturbance had a significant
effect on H´ for day 97, both when POC:PP was included
in the model (P = 0.016) and when it was not (P = 0.005)
(Table IIId,e). This indicates that the effect of disturbance on species diversity was not the result of correlated
resource limitation. We found similar results when including
POC:PN in the model (P = 0.015 and P = 0.005 with and
without POC:PN in the model, respectively). The Dunnett’s
post hoc comparison showed that the 7-d disturbance level
had a significantly lower H´ than the undisturbed treatment
(P = 0.006). Furthermore, there was no significant direct
correlation between POC:PP and H´ (r = 0.63, P = 0.067).
The changes in diversity over time depended both on
the disturbance level and on the community composition
(Figure 3). H’ increased for all treatments between the
inoculation of the microcosms (t = 0) and the first sampling
at 14 d. After 14 d, however, H’ declined for all microcosms
exposed to the 7-d and 11-d disturbance levels, although the
433
NARWANI, BERTHIN & MAZUMDER: ENDOGENOUS DYNAMICS MAINTAIN SPECIES DIVERSITY
FIGURE 3. Shannon Diversity Index (H´) as a function of time for individual
replicates. Each panel shows both replicates of all 3 phytoplankton community compositions (CCs). CC1 is represented by circles, CC2 is represented
by triangles, and CC3 is represented by squares. The open and closed symbols signify separate replicates of the same treatment. The thick solid line
represents the mean H´ across all treatments and replicates at a given time.
a) 7-d disturbance interval b) 11-d disturbance interval c) undisturbed.
434
decline was less rapid for the 11-d disturbance treatment.
Changes in diversity in the undisturbed microcosms after
14 d were more variable, with some community composition replicates increasing in diversity, others remaining
stable, and yet others declining (Figure 3).
In all of the compositions receiving the 7-d disturbance treatment the distribution of relative species abundance changed during the first month of the experiment
(Figure 4). At day 97, however, Chlamydomonas reinhardtii
accounted for 95.0% or more of the total biomass in each
microcosm. Species relative abundance distributions varied
over a longer period of time in some of the microcosms
exposed to the 11-d disturbance treatment (Figure 4). For
composition 1, C. reinhardtii established itself as the dominant after 14 d and accounted for more than 94% of the
total biomass at 97 d in both replicates, much like the
dynamics for this composition under the 7-d disturbance
interval. In compositions 2 and 3, the relative abundance of
each species fluctuated over a longer period of time, with
Cryptomonas cf. rostratiformis becoming established as
the dominant for both compositions at 97 d. For these compositions, Ankistrodesmus falcatus also accounted for an
average of at least 12% of the biomass at the termination of
the experiment, indicating that C. cf. rostratiformis had not
completely excluded this species. The relative abundances
of species in the undisturbed microcosms continued changing
throughout the experiment for all compositions. Competitive
exclusion had not occurred in any undisturbed microcosm
after 97 d (Figure 4).
Species that were not included in the inoculated community compositions occurred in 1 or more of the replicates
of each treatment by the final sampling date (Appendix I;
Figure 4). Cyclotella sp. was the most common invader, followed by C. reinhardtii and A. falcatus. Chlamydomonas.
reinhardtii only invaded microcosms receiving the 7-d disturbance treatment, while A. falcatus invaded microcosms
receiving both the 11-d disturbance and undisturbed treatments. An unidentified species occurred in a single sample
from an undisturbed microcosm with community composition 3 (Figure 4). The invasions of the 7-d disturbance treatments by C. reinhardtii caused the community composition
and the disturbance treatments to become conflated because
all of the 7-d disturbance treatments became dominated by
the same species by the end of the experiment (Figure 4).
Therefore, in order to separate the effect of disturbance
frequency and community composition, we compared the
diversity among disturbance treatments within a community
composition in which C. reinhartdtii was present from the
time of inoculation (namely, community composition 1). A
one-way ANOVA to detect the effect of disturbance within
community composition 1 showed that there was a significant effect of disturbance (F = 108.582, P = 0.002, df = 2,
Figure 4). A Dunnett’s post hoc comparison also showed
that the undisturbed treatment had a significantly higher
diversity at day 97 than either of the disturbed treatments
(P = 0.002 for both comparisons). This indicates that the
presence of C. reinhardtii was not the reason for the low
diversity found in the 7-d disturbance treatments, because
diversity was maintained in the undisturbed microcosms of
composition 1 despite the fact that they contained this species.
ÉCOSCIENCE, VOL. 16 (4), 2009
Disturbance had a significant effect on ln-transformed
particulate organic carbon (ln-POC) (F = 1094.460,
P < 0.001, df = 3, Appendix IIa). The ln-POC of the undisturbed treatment was significantly different from the 7-d
and 11-d disturbance treatments (P < 0.05). The undisturbed
treatment also had the greatest proportion of particulate
nitrogen (Appendix IIb). By contrast, both the 7-d and
undisturbed treatments had comparably high particulate
phosphorus levels, while the 11-d disturbance treatment had
the lowest level of particulate phosphorus (Appendix IIc).
FIGURE 4. Species-specific relative biovolume under the 7-d, 11-d, and undisturbed treatments, respectively from left to right. Community composition
CC1 is in the top panel, CC2 is in the middle, and CC3 is on the bottom. An asterisk (*) signifies an invading species (Appendix I). Relative biovolume is
averaged across the 2 replicates for each community composition × disturbance treatment cross (see Table II).
435
NARWANI, BERTHIN & MAZUMDER: ENDOGENOUS DYNAMICS MAINTAIN SPECIES DIVERSITY
Disturbance also had a significant effect on the particulate
C:N ratios (F = 5.377, P = 0.039, df = 3) and ln(C:P) ratios
(F = 287.774, P < 0.001, df = 3). Specifically, the average
particulate C:N and C:P molar ratios were both lowest for
the 7-d disturbance treatment and greatest for the undisturbed treatment (Table II).
Discussion
Our results demonstrate that greater diversity can be
maintained in undisturbed assemblages of phytoplankton than in assemblages experiencing periodic, externally
forced fluctuations in density and nutrient availability. This
indicates that, at least in our experimental system, endogenously generated competitive dynamics were more effective
in maintaining high species diversity than exogenously
generated variability in resource availability. Community
composition and disturbance also had significant interactive effects on diversity. While the greatest diversity
always occurred in undisturbed microcosms, the form of the
increase in diversity from the 7-d disturbance treatment to
the undisturbed treatment depended on the identity of species within each assemblage (Figure 2).
Disturbances are generally considered to be discrete
events that result in the loss of biomass or liberation of
resources (Grime, 1977). Many studies have tested the
importance of disturbance on maintaining species diversity,
and 2 prominent hypotheses have been proposed regarding
the relationships between disturbance and diversity. The
intermediate disturbance hypothesis (IDH) predicts a unimodal relationship between disturbance and diversity
because intermediate frequencies of disturbance revert communities to intermediate states of succession (Grime, 1973;
Connell, 1978; Hastings, 1980). The IDH has been supported by some experimental tests both in the lab (Robinson
& Sandgren, 1983; Gaedeke & Sommer, 1986; Sommer,
1995; Flöder, Urabe & Kawabata, 2002) and in the field
(Flöder & Sommer, 1999). The second hypothesis, known
as the dynamic equilibrium model, proposes that diversity
is determined by the interaction between growth rates and
the frequency of disturbance (Huston, 1979). The highest
diversity is predicted to occur at low to moderate growth
rates and frequencies of disturbance. The diversity–disturbance relationship is also predicted to depend on productivity because this may affect growth rates (Kondoh,
2001). Currently, there is some experimental support for this
hypothesis (Proulx et al., 1996; Proulx & Mazumder, 1998;
Worm et al., 2002; but see Scholes, Warren & Beckerman,
2005). Disturbances do not promote coexistence by
reducing the strength of competition (Chesson & Huntly,
1997). Instead, they create temporal environmental heterogeneity, acting as limiting factors or niches along which
species can differentiate (Chesson & Huntly, 1997; Chase &
Leibold, 2003). In producing heterogeneity in limiting factors, disturbances can allow the expression of storage effects
or competitive relative nonlinearities, which enable coexistence (Koch, 1974a; Chesson, 2000a).
Much of the ecological literature on diversity–disturbance relationships postulates a peaked relationship (Connell,
1978; Huston, 1979; Mackey & Currie, 2001). However, the
436
full gamut of possible disturbance–diversity relationships
has been observed (Mackey & Currie, 2001), including
monotonically decreasing (Haddad et al., 2008), increasing (Sommer, 1993), and U-shaped (Flöder & Burns, 2004)
relationships. A positive relationship might be detected
when the upper end of a peaked relationship is not sampled;
i.e., the frequency of disturbance can increase infinitely, and
so the highest frequencies of disturbance are not well represented. Conversely, a negative relationship could be generated
by neutral dynamics (Hubbell, 2001). In this case, disturbance reduces the population density of species and could
lead to a greater number of extinctions due to demographic
stochasticity. A negative relationship could also be detected
if internal dynamics, including intransitive and nonlinear
competition, more effectively maintained species diversity
than disturbance. In this case, disturbance would effectively
disrupt the internal dynamics that maintain diversity. We
observed such a negative relationship in our study (note:
our figures have disturbance interval and not frequency on
the x-axis). We are unaware of any explicit mechanism that
can generate a U-shaped relationship. This rarely-observed
relationship may be the outcome of more than 1 mechanism
acting at once.
Our results suggest that the internal competitive
dynamics occurring within 3 different phytoplankton assemblages were more effective at maintaining species diversity
than the environmental temporal heterogeneity in resource
availability produced by 2 different frequencies of disturbance (Figure 2). Other recent studies have also found relatively long-term coexistence of multiple species in purely
competitive communities (e.g., Kerr et al., 2002; Roelke,
Augustine & Buyukates, 2003; Haddad et al., 2008),
although they did not compare the importance of internal
versus external dynamics per se. There are at least 3 potential mechanisms that could lead to internally mediated competitive coexistence. The first 2 mechanisms require either
that the phytoplankton in our experiment were competing
for more than 3 abiotic resources and display relative nonlinearity in competitive abilities for these limiting factors
(Armstrong & McGehee, 1980; Huisman & Weissing, 1999;
2001) or that they are ordered in an intransitive competitive
hierarchy or network (Huismann & Weissing, 2001; Kerr
et al., 2002). There has as yet been no experimental demonstration that relative nonlinearity in competitive abilities
promotes coexistence in internally regulated assemblages,
and this remains an important avenue for future research.
By contrast, however, Kerr et al. (2002) experimentally
demonstrated that when bacteria compete locally in a game
of “rock–paper–scissors” (competitive intransitivity), a
greater number of species than the number of limiting
resources may coexist. The third mechanism is the limitation of the community by as many limiting factors as there
are species (Tilman, 1982).
The undisturbed microcosms in our experiment were
effectively closed systems (i.e., dissolved nutrients were
not replaced from an exterior source), and it is therefore
likely that they were limited by a larger number of resources
than the disturbed microcosms. Indeed, the nutrient analysis showed that the phytoplankton were more limited with
respect to nitrogen and phosphorus in the undisturbed than
ÉCOSCIENCE, VOL. 16 (4), 2009
the disturbed microcosms (Table II; Appendix II). However,
when the level of nutrient limitation was added into the
ANOVA as a covariate, it was not a significant explanatory
variable, and the effect of disturbance remained significant, indicating that the importance of disturbance was not
mediated by its effect on nutrient limitation (Table III). The
level of nutrient limitation did not explain the higher species diversity in the undisturbed microcosms. Furthermore,
Tilman’s (1982) predictions are relevant to communities
at equilibrium, and the only communities that appeared to
have reached equilibrium were those that were disturbed
most frequently (Figure 3), whereas the undisturbed communities continued to vary in relative species abundances
until day 97. Such non-equilibrium dynamics are more
similar to those predicted in cases of relatively nonlinear
competition and intransitivity (Huisman & Weissing, 2001;
Roelke, Augustine & Buyukates, 2003). It is possible, however, that because the microcosms were closed systems, the
identity of limiting resources varied over time as different
species came to dominate. This, in turn, may have prevented
equilibrium in a system that would have approached equilibrium if nutrients were continuously replenished. We cannot
discriminate between these possibilities further here, but
what is clear from our study is that internal dynamics more
effectively maintained species diversity than exogenous
variability. Direct testing to tease apart the importance of
different internal mechanisms would be a valuable avenue
for future research.
Depending on the disturbance regime, some compositions were less likely to be dominated by a single species
over time than others. This stands in direct contrast to
predictions from neutral models that predict that species
traits are unimportant in determining community diversity
or coexistence (Hubbell, 2001). A recent grassland experiment showed that the ability of communities to maintain
species diversity in a disturbed landscape depended on
the degree of functional complementarity among species
(Questad & Foster, 2008). Also, microcosm experiments
have shown that species’ intrinsic growth rates are important in determining responses to disturbance (Haddad et al.,
2008). Niche theory predicts that species characteristics
such as their requirements for, and impacts upon, resources
are important in determining community composition via
species sorting (Chase & Leibold, 2003). Also, Huisman
and Weissing (2001) predicted that supersaturated coexistence would be more likely to occur in communities where
species showed competitive intransitivity than in those that
did not. Furthermore, the storage effect and relative nonlinearity require particular differences in species traits in
order to promote coexistence (Chesson, 2000b). Numerous
theoretical approaches therefore predict the importance of
species traits and hence community composition on coexistence that we have demonstrated here.
The effect of community composition in determining
coexistence depended on the level of disturbance of the
microcosms. In the 7-d disturbance treatments, regardless
of the community composition, the microcosms came to be
dominated by a single species, C. reinhardtii, whereas in the
undisturbed microcosms there was greater variability in the
final outcome of competitive dynamics, which depended on
the original community composition. This indicates that disturbance was a strong homogenizing force among communities, causing “community convergence” sensu Houseman
et al. (2008). Similar homogenizing effects of disturbance
have been found in grassland (Collins & Smith, 2006) and
pond communities (Chase, 2003). The alternative interpretation of the significant interaction between composition and
disturbance is that the effects of disturbance depended on
the community composition. For instance, in the 7-d and
11-d disturbance treatments, fast-growing species more
quickly dominated communities (Table 1; Figure 3), whereas slower-growing species, and potentially also better competitors for limiting nutrients, were better able to persist in
the undisturbed microcosms. This agrees with the results
of Haddad et al. (2008), who showed that the persistence of
species in the face of disturbance could be predicted most
effectively by their intrinsic rates of increase.
The occurrence of invasions indicates that even the
undisturbed microcosms were not completely “closed”
to propagules of new species. Nevertheless, because the
invaders were competitors (and not predators, parasites, or
resources), they fit into the class of organisms that by definition take part in endogenous dynamics. As a result, the
invasions did not prevent us from asking whether endogenous or exogenous mechanisms are more important for
maintaining species diversity. We are confident that the
propagule pressure was randomized among microcosms
because the microcosms were randomized in space (in the
culture chamber) and in time (during sampling). Unbiased
propagule pressure should have affected diversity equally
across all of the treatments, if at all. We observed that,
with only 1 exception, microcosms experienced no more
than 1 invasion, and so there was no evidence that the invasions added species to communities differentially across
treatments (Appendix I). Even if the invasions had been
concomitant with some indiscernible external source of
structure or variation upon which storage dynamics and
nonlinearities could be expressed, we would expect the
diversity among the disturbance treatments to become more
similar. As a result, the difference that we observed between
disturbance treatments was conservative. Our data show that
despite the fact that invasions occurred across the board, the
undisturbed treatments were still significantly more diverse.
When invaders were able to establish in the 7-d disturbance
treatments, they either remained at low density or became
dominant (Figure 4). In the undisturbed microcosms, invading species did not become dominant but rather became a
part of relatively even communities containing significant
fractions of biovolume of more than 1 species.
With respect to the effect of community composition
on diversity, we expected that invasions would reduce differences in community composition, thereby diminishing
any effect of composition on diversity. As a result, a significant difference in diversity among compositions would
be conservative. Disturbance frequency may have selected
for the establishment of particular invader species (Collins
& Smith, 2006; Houseman et al., 2008). This could have
caused the communities to become more similar within a
given level of disturbance (Figure 4). The 7-d disturbance
treatment allowed only C. reinhardtii to invade, and this
437
NARWANI, BERTHIN & MAZUMDER: ENDOGENOUS DYNAMICS MAINTAIN SPECIES DIVERSITY
species became dominant in all of the microcosms. This
reduced the original differences in community composition and likely prevented the detection of any effect of the
original community composition on diversity. Community
convergence in the 7-d treatments also caused community composition and the disturbance treatment to be conflated. Nevertheless, when we controlled for the invasion of
C. reinhardtii by looking only at the communities where this
species was always present (composition 1), we found that
disturbance still had a significant effect on diversity. In the
undisturbed microcosms, invasions did not lead to monodominance and community convergence, and so we were
able to detect the effect of community composition on species diversity. The effect of community composition in the
undisturbed microcosms is conservative and was robust to
the effect of the invasions. Together these results may explain
the interaction between composition and disturbance.
In conclusion, we have shown that competition for
resources among phytoplankton species can enable greater
coexistence in closed systems than in systems experiencing
externally generated variability in resource availability over
time. We recommend that greater attention be paid to the
potential for internal competitive dynamics to maintain species diversity and that similar studies be conducted in other
experimental systems to determine the generality of our
findings. Historically, the emphasis in biodiversity studies
has been placed on the importance of external forces in generating long-term coexistence, but theory and our empirical
work now suggest that internal dynamics may be equally
(if not more) important in maintaining biodiversity in some
communities. Coexistence in our experimental communities
also depended on the interaction between the composition
and disturbance level. This confirmed the predicted importance of species traits in maintaining biodiversity in our
experimental communities.
Acknowledgements
We thank A. Tokarek, L. Laverty, and S. Bozukova for their
technical assistance and P. Kratina and U. Narwani for their helpful comments on the manuscript. We also thank one anonymous
reviewer and S. Roy for comments that improved the manuscript. This
research was supported by an Natural Sciences and Engineering
Research Council of Canada (NSERC) CGSM to A. Narwani, an
NSERC USRA to J. Berthin, and NSERC DG and IRC grants to
A. Mazumder.
Literature cited
Abrams, P., 2001. The effect of density-independent mortality on
the coexistence of exploitative competitors for renewing resources. American Naturalist, 158: 459–470.
Abrams, P. & R. Holt, 2002. The impact of consumer-resource
cycles on the coexistence of competing consumers. Theoretical
Population Biology, 62: 281–295.
Armstrong, R. A. & R. McGehee, 1980. Competitive exclusion.
American Naturalist, 115: 151–170.
Buss, L. W. & J. B. C. Jackson, 1979. Competitive networks:
Nontransitive competitive relationships in cryptic coral reef
environments. American Naturalist, 113: 223–234.
Caswell, H., 1978. Predator-mediated coexistence: A nonequilibrium model. American Naturalist, 112: 127–154.
438
Chase, J. M., 2003. Community assembly: When should history
matter? Oecologia, 136: 489–498.
Chase, J. M. & M. A. Leibold (eds.), 2003. Ecological Niches:
Linking Classical and Contemporary Approaches. University of
Chicago Press, Chicago, Illinois.
Chesson, P., 2000a. General theory of competitive coexistence
in spatially-varying environments. Theoretical Population
Biology, 58: 211–237.
Chesson, P., 2000b. Mechanisms of maintenance of species diversity. Annual Review of Ecology and Systematics, 31: 343–366.
Chesson, P. & N. Huntly, 1997. The roles of harsh and fluctuating conditions in the dynamics of ecological communities.
American Naturalist, 150: 519–553.
Chesson, P. L. & R. R. Warner, 1981. Environmental variability
promotes coexistence in lottery competitive systems. American
Naturalist, 117: 923–943.
Clesceri, L. S., A. E. Greenberg & A. D. Eaton, 1998. Standard
Methods for the Examination of Water and Wastewater. 20
Edition. American Public Health Association, Washington, DC.
Collins, S. L. & M. D. Smith, 2006. Scale-dependent interaction of
fire and grazing on community heterogeneity in tallgrass prairie. Ecology, 87: 2058–2067.
Connell, J. H., 1978. Diversity in tropical rain forests and coral
reefs. Science, 199: 1302–1310.
Crowley, P. H., 1979. Predator-mediated coexistence: An equilibrium interpretation. Journal of Theoretical Biology, 80: 129–144.
DeAngelis, D. L. & J. C. Waterhouse, 1987. Equilibrium and
nonequilibrium concepts in ecological models. Ecological
Monographs, 57: 1–21.
Descamps-Julien, B. & A. Gonzalez, 2005. Stable coexistence in
a fluctuating environment: An experimental demonstration.
Ecology, 86: 2815–2824.
Flöder, S. & C. Burns, 2004. Phytoplankton diversity of shallow
tidal lakes: Influence of period salinity changes on diversity and species number of a natural assemblage. Journal of
Phycology, 40: 54–61.
Flöder, S. & U. Sommer, 1999. Diversity in planktonic communities: An experimental test of the intermediate disturbance
hypothesis. Limnology and Oceanography, 44: 1114–1119.
Flöder, S., J. Urabe & Z. Kawabata, 2002. The influence of fluctuating light intensities on species composition and diversity of
natural phytoplankton communities. Oecologia, 133: 395–401.
Gaedeke, A. & U. Sommer, 1986. The influence of the frequency
of periodic disturbances on the maintenance of phytoplankton
diversity. Oecologia, 71: 25–28.
Gause, G. F., 1934. The Struggle for Existence. Hafner Publishing
Company, New York, New York.
Grime, J. P., 1973. Competitive exclusion in herbaceous vegetation. Nature, 242: 344–347.
Grime, J. P., 1977. Evidence for the existence of three primary
strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist, 111: 1169–1194.
Haddad, N. M., M. Holyoak, T. M. Mata, K. F. Davies, B. A.
Melbourne & K. Preston, 2008. Species’ traits predict the
effects of disturbance and productivity on diversity. Ecology
Letters, 11: 348–356.
Hardin, G., 1960. The competitive exclusion principle. Science,
131: 1292–1297.
Hastings, A., 1980. Disturbance, coexistence, history and competition for space. Theoretical Population Biology, 18: 363–373.
Hecky, R. E. & P. Kilham, 1988. Nutrient limitation of phytoplankton in freshwater and marine environments: A review of
recent evidence on the effects of enrichment. Limnology and
Oceanography, 33: 796–822.
ÉCOSCIENCE, VOL. 16 (4), 2009
Hillebrand, H., C. D. Durselen, D. Kirschtel, U. Pollinger & T.
Zohary, 1999. Biovolume calculation for pelagic and benthic
microalgae. Journal of Phycology, 35: 403–424.
Holt, R. D., 1984. Spatial heterogeneity, indirect interactions, and
the coexistence of prey species. The American Naturalist, 124:
377–406.
Houseman, G. R., G. G. Mittlebach, H. L. Reynolds & K. L.
Gross, 2008. Perturbations alter community convergence, divergence, and formation of multiple community states. Ecology,
89: 2172–2180.
Hubbell, S. P., 2001. The Unified Neutral Theory of Biodiversity
and Biogeography. Monographs in Population Biology 32.
Princeton University Press, Princeton, New Jersey.
Huisman, J. & F. J. Weissing, 1999. Biodiversity of plankton by
species oscillations and chaos. Nature, 402: 407–410.
Huisman, J. & F. J. Weissing, 2001. Biological conditions for
oscillations and chaos generated by multispecies competition.
Ecology, 82: 2682–2695.
Huston, M., 1979. A general hypothesis of species diversity.
American Naturalist, 113: 81–101.
Hutchinson, G. E., 1961. The paradox of the plankton. American
Naturalist, 95: 137–145.
Janzen, D. H., 1970. Herbivores and the number of tree species in
tropical forests. American Naturalist, 104: 501–528.
Kelly, C. & M. G. Bowler, 2002. Coexistence and relative abundance in forest trees. Nature, 417: 437–440.
Kerr, B., M. A. Riley, M. W. Feldman & B. J. M. Bohannan, 2002.
Local dispersal promotes biodiversity in a real-life game of
rock–paper–scissors. Nature, 418: 171–174.
Kilham, S. S., D. A. Kreegar, S. G. Lynn, C. E. Goulden & L.
Herrara, 1998. COMBO: A defined freshwater culture medium
for algae and zooplankton. Hydrobiologia, 337: 147–159.
Kirkup, B. C. & M. A. Riley, 2004. Antibiotic-mediated antagonism leads to bacterial game of rock–paper–scissors in vivo.
Nature, 428: 412–414.
Koch, A. L., 1974a. Coexistence resulting from an alternation of
density dependent and density independent growth. Journal of
Theoretical Biology, 44: 373–386.
Koch, A. L., 1974b. Competitive coexistence of two predators utilizing the same prey under constant environmental conditions.
Journal of Theoretical Biology, 44: 387–395.
Kondoh, M., 2001. Unifying the relationships of species richness to productivity and disturbance. Proceedings of the Royal
Society of London B, 268: 269–271.
Levin, S. A., 1970. Community equilibria and stability, and an
extension of the competitive exclusion principle. American
Naturalist, 104: 413–423.
Levins, R. & D. Culver, 1971. Regional coexistence of species and
competition between rare species. Proceedings of the National
Academy of Sciences of the United States of America, 68:
1246–1248.
MacArthur, R., 1969. Species packing, and what interspecies competition minimizes. Proceedings of the National Academy of
Sciences of the United States of America, 64: 1369–1371.
MacArthur, R. & R. Levins, 1964. Competition, habitat selection, and character displacement in a patchy environment.
Proceedings of the National Academy of Sciences of the United
States of America, 51: 1207–1210.
MacArthur, R. & R. Levins, 1967. The limiting similarity, convergence, and divergence of coexisting species. American
Naturalist, 101: 377–385.
Mackey, R. L. & D. J. Currie, 2001. The diversity–disturbance
relationship: Is it generally strong and peaked? Ecology, 82:
3479–3492.
Proulx, M. & A. Mazumder, 1998. Reversal of grazing impact on
plant species richness in nutrient-poor versus nutrient-rich ecosystems. Ecology, 79: 2581–2592.
Proulx, M., F. R. Pick, A. Mazumder, P. B. Hamilton & D. R. S.
Lean, 1996. Experimental evidence for interactive impacts
of human activities on lake algal species richness. Oikos, 76:
191–195.
Questad, E. J. & B. L. Foster, 2008. Coexistence through spatiotemporal heterogeneity and species sorting in a grassland plant
community. Ecology Letters, 11: 717–726.
Quinn, G. P. & M. J. Keough, 2002. Experimental Design and
Data Analysis for Biologists. Cambridge University Press,
Cambridge.
Rescigno, A. & I. W. Richardson, 1965. On the competitive exclusion principle. Bulletin of Mathematical Biophysics, 27: 85–89.
Robinson, J. V. & C. D. Sandgren, 1983. The effect of temporal
environmental heterogeneity on community structure: A replicated experimental study. Oecologia, 57: 98–102.
Roelke, D., S. Augustine & Y. Buyukates, 2003. Fundamental predictability in multispecies competition: The influence of large
disturbance. American Naturalist, 162: 615–623.
Scheffer, M. & E. H. Van Nes, 2006. Self-organized similarity, the evolutionary emergence of groups of similar species.
Proceedings of the National Academy of Sciences of the United
States of America, 103: 6230–6235.
Schippers, P., A. M. Verschoor, M. Vos & W. M. Mooij, 2001. Does
“supersaturated coexistence” resolve the “paradox of the plankton”? Ecology letters, 4: 404–407.
Scholes, L., P. H. Warren & A. P. Beckerman, 2005. The combined
effects of energy and disturbance on species richness in protist
microcosms. Ecology Letters, 8: 730–738.
Shannon, C. E. & W. Weaver, 1949. The Mathematical Theory of
Communication. University of Illinois Press, Champaign, Illinois.
Sinervo, B. & C. M. Lively, 1994. The rock–paper–scissors game
and the evolution of alternative male strategies. Nature, 380:
240–243.
Sommer, U., 1995. An experimental test of the intermediate disturbance hypothesis using cultures of marine phytoplankton.
Limnology and Oceanography, 40: 1271–1277.
Steiner, C. F., 2005. Impacts of density-independent mortality
and productivity on the strength and outcome of competition.
Ecology, 86: 727–739.
Tilman, D., 1977. Resource competition between planktonic
algae: An experimental and theoretical approach. Ecology, 58:
338–348.
Tilman, D., 1981. Tests of resource competition theory using four
species of Lake Michigan algae. Ecology, 62: 802–815.
Tilman, D., 1982. Resource Competition and Community
Structure. Princeton University Press, Princeton, New Jersey.
Tilman, D., 1994. Competition and biodiversity in spatially structured habitats. Ecology, 75: 2–16.
Underwood, A. J., 1997. Experiments in Ecology. Cambridge
University Press, Cambridge.
Volterra, V., 1928. Variations and fluctuations of the number of
individuals in animal species living together. Journal du Conseil
Permanent International pour l’Exploration de la Mer, 3: 3–51.
Worm, B., H. K. Lotze, H. Hillebrand & U. Sommer, 2002.
Consumer versus resource control of species diversity and ecosystem functioning. Nature, 417: 848–851.
439
NARWANI, BERTHIN & MAZUMDER: ENDOGENOUS DYNAMICS MAINTAIN SPECIES DIVERSITY
APPENDIX I. Replicate numbers and the species that invaded them.
Each species initially appeared at various sampling dates after
14 d.
Replicate
Disturbance
treatment
Community
composition
1
2
3
4
5
6
7
8
9
10
11
12
13
13
14
15
16
17
7-d
7-d
7-d
7-d
7-d
7-d
11-d
11-d
11-d
11-d
11-d
11-d
Undisturbed
Undisturbed
Undisturbed
Undisturbed
Undisturbed
Undisturbed
1
1
2
2
3
3
1
1
2
2
3
3
1
1
1
2
2
3
440
Contaminant
Cyclotella sp.
Cyclotella sp.
Chlamydomonas reinhardtii
Chlamydomonas reinhardtii
Chlamydomonas reinhardtii
Chlamydomonas reinhardtii
Cyclotella sp.
Cyclotella sp.
Cyclotella sp.
Cyclotella sp.
Ankistrodesmus falcatus
Ankistrodesmus falcatus
Cyclotella sp.
Unknown species
Cyclotella sp.
Cyclotella sp.
Cyclotella sp.
Ankistrodesmus falcatus
APPENDIX II. Nutrient concentrations (μmol·L–1) as sampled from
the microcosms on day 97. Grey bars represent the particulate fraction and black bars represent the dissolved fraction. Values for each
bar represent the mean of 2 replicates for each community composition × disturbance treatment cross. Community compositions
increase from left to right in each panel, starting at community
composition 1. Error bars represent + SE of the total concentration.
a) Organic carbon b) Nitrogen c) Phosphorus.